Description of problem: [Mon Mar 26 11:05:14 2007] [error] (9)Bad file descriptor: apr_socket_accept: (client socket) Version-Release number of selected component (if applicable): httpd-2.2.3-6.el5 How reproducible: Happens everytime I run apachectl graceful Steps to Reproduce: 1. apachectl graceful 2. and again apachectl graceful etc. etc. 3. Actual results: Bad file descriptor: apr_socket_accept: (client socket) Expected results: No error Additional info:
Created attachment 150890 [details] strace -f apachectl graceful >& strace-httpd.txt
Thanks for the report. This is actually a harmless error; it occurs because the socket on which the server calls accept() is immediately closed in child processes, upon receipt of the graceful restart signal. It may be possible to safely hide it from the logs or avoid it somehow, however.
According to a Debian website, this problem has been isolated and fixed in release 2.2.6-2 from the Apache team. All that would be required of the Fedora project is to repackage the stuff into an FC7 (and FC8) RPM. Who is the right person to take this on?
1. This was filled against RHEL 2. did httpd get patched with this fix 3. Since this is harmless and just an noise it can be closed as FIXED
There was a set of changes related to graceful restart handling proposed by a Debian developer, but I don't think it will have made a difference for the single-listener case as is covered here. It does look like it's possible to avoid logging this in the graceful restart case.
created RHTS test /CoreOS/httpd/misc/bug233955
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2009-1380.html